Benchmarking dual-level MS-Tor and DLPNO-CCSD(T) methods for H-abstraction from methyl pentanoate by an OH radical

Methyl pentanoate (MP) was recently proposed as a potential biodiesel surrogate due to its negative temperature coefficient region. To provide a basis for constructing an accurate mechanism, chemical kinetics of H-abstraction from MP by an OH radical are investigated theoretically at 200-2000 K. M06-2X/cc-pVTZ is applied for geometry optimizations and frequency calculations. Given the long alkyl-chain in MP, the multi-structural torsional anharmonicity is characterized by using the dual-level MS-T method due to its relatively low computational cost and established accuracy. In particular, AM1 and M06-2X/cc-pVTZ are adopted as low-level and high-level methods in dual-level MS-T, respectively. The results of dual-level MS-T are further used to benchmark against the full high-level method (M06-2X/cc-pVTZ), leading to an uncertainty of less than 30% in the high temperature range. For the single-point energy calculations, the lower computational cost DLPNO-CCSD(T) method is first used to benchmark against the gold-standard method CCSD(T) for small methyl ester (C₂-C₄) + OH reaction systems, yielding an overestimation of less than 1.1 kcal mol^-1 for barrier height; it is then used to refine the electronic energies for the present reaction system MP + OH. Phase-space theory and conventional transition state theory are used to calculate the H-abstraction rate constants. After compensating the uncertainty of barrier height, the calculated phenomenological H-abstraction rate constants agree well with the experimental data at 263-372 K. Branching ratio analysis indicates that the β-site H abstraction is the dominant channel at 200-1200 K due to its lowest barrier height.